EX 4883

Three theoretical models have been developed for the prediction of vasodilatory activity through a series of 2‐D molecular descriptors. A database of 501 compounds was selected from the literature and, of these compounds, 86 have vasodilatory activity. The QSAR models are capable of differentiating between active and inactive compounds with a level of classification greater than 80%. The models were used to predict the activity of a series of coumarins derived from resveratrol (a natural compound that is present in wine and has good vasodilatory activity) and led to the synthesis of three selected molecules. The sysnthesis of the resveratrol‐coumarin hybrids was efficiently achieved through a straightforward and direct route, and their vasodilatory activities were determined experimentally in rat aorta rings that were precontreated with noradrenaline. The theoretical results (“in silico” evaluation) show very good agreement with the experimental data (“in vitro” evaluation), which provide evidence of the reliability of the theoretical calculations and show their applicability in the rational design of new compounds. The compound predicted by the three models to be active (compound6) was confirmed to be the more active thantrans‐resveratrol.